Está en la página 1de 23

Estudios de Psicologa

Studies in Psychology

ISSN: 0210-9395 (Print) 1579-3699 (Online) Journal homepage: http://www.tandfonline.com/loi/redp20

Comparative Psychotherapy Process Scale


(CPPS): analysis of the validity and reliability of
its adaptation into Spanish / Escala Comparativa
del Proceso Psicoteraputico (CPPS): anlisis
de la validez y confiablidad de su adaptacin al
castellano

Juan-Martn Gmez-Penedo, Juan-Carlos Barrientos, Ariana Martnez &


Andrs Roussos

To cite this article: Juan-Martn Gmez-Penedo, Juan-Carlos Barrientos, Ariana Martnez


& Andrs Roussos (2016): Comparative Psychotherapy Process Scale (CPPS): analysis of
the validity and reliability of its adaptation into Spanish / Escala Comparativa del Proceso
Psicoteraputico (CPPS): anlisis de la validez y confiablidad de su adaptacin al castellano,
Estudios de Psicologa, DOI: 10.1080/02109395.2016.1189207

To link to this article: http://dx.doi.org/10.1080/02109395.2016.1189207

Published online: 29 Jul 2016.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at


http://www.tandfonline.com/action/journalInformation?journalCode=redp20

Download by: [Juan Martin Gomez Penedo] Date: 29 July 2016, At: 23:15
Estudios de Psicologa / Studies in Psychology, 2016
http://dx.doi.org/10.1080/02109395.2016.1189207

Comparative Psychotherapy Process Scale (CPPS): analysis of


the validity and reliability of its adaptation into Spanish / Escala
Comparativa del Proceso Psicoteraputico (CPPS): anlisis de la
validez y confiablidad de su adaptacin al castellano
Juan-Martn Gmez-Penedoa, Juan-Carlos Barrientosb, Ariana Martnezc,
and Andrs Roussosa
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

a
Universidad de Buenos Aires and CONICET; bUniversidad de San Francisco Xavier de
Chuquisaca; cUniversidad de Belgrano
(Received 11 August 2015; accepted 29 April 2016)

Abstract: The study of psychotherapeutic interventions and their link with


process and outcome variables has become a popular topic of great interest in
psychotherapy research. This paper presents the adaptation into Spanish of the
Comparative Psychotherapy Process Scale (CPPS), a 20-item scale that mea-
sures the degree to which therapists use Cognitive-Behavioural or
Psychoanalytic-Interpersonal interventions. A hundred and one therapists
(50 cognitive-behavioural and 51 psychoanalytic) completed the CPPS for
the last patient they had treated. In a second sample of 15 patient-therapist
dyads, both participants completed the CPPS regarding the same session. The
subsequent analyses showed preliminary evidence of reliability (internal con-
sistency and homogeneity) and validity (of construct and criterion) in the
adaptation of the CPPS to Spanish. The implications of having a scale of these
characteristics in Spanish and future lines of research are discussed.
Keywords: Comparison Psychotherapy Process Scale; CPPS; validity;
reliability; adaptation

Resumen: El estudio de las intervenciones psicoteraputicas y su vnculo con


variables de proceso y resultado se ha convertido en uno de los tpicos de
mayor inters en la investigacin en psicoterapia. Este trabajo presenta la
adaptacin al castellano de la Escala Comparativa del Proceso
Psicoteraputico (CPPS), una escala de 20 tems que mide el grado en que
terapeutas utilizan intervenciones Cognitivo-Conductuales o Psicoanalticas-
Interpersonales. 101 terapeutas (50 cognitivos-conductuales y 51
psicoanalticos) completaron la CPPS respecto del ltimo paciente que
haban atendido. En una segunda muestra de 15 dadas paciente-terapeuta,
ambos participantes completaron la CPPS respecto de la misma sesin. Los

English version: pp. 110 / Versin en espaol: pp. 1120


References / Referencias: pp. 2022
Translated from Spanish / Traduccin del espaol: Liza DArcy
Authors Address / Correspondencia con los autores: Juan-Martn Gmez-Penedo,
Universidad de Buenos Aires y CONICET, Junn 1137 Piso 5 Departamento A, Buenos
Aires, Argentina. E-mail: juanmartin.gomez@comunidad.ub.edu.ar

2016 Fundacion Infancia y Aprendizaje


2 J.-M. Gmez-Penedo et al.

anlisis conducidos muestran evidencias preliminares de la confiabilidad


(consistencia interna y homogeneidad) y validez (de constructo y de criterio)
en la adaptacin de la CPPS. Las implicancias de contar con una escala de
estas caractersticas en castellano y futuras lneas de investigacin son
discutidas.
Palabras clave: Escala Comparativa del Proceso Psicoteraputico; CPPS;
validez; confiablidad; adaptacin

There is a strong presumption in clinicians, academics and researchers that


interventions the verbal actions that therapists perform during sessions aimed
at patients (Roussos, Etchebarne, & Waizmann, 2005) are one of the main
mechanisms of change in psychotherapy (McCarthy & Barber, 2009). This
assumption has found partial support through empirical research in different
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

theoretical frameworks (Owen & Hilsenroth, 2011; Strunk, Brotman, &


DeRubeis, 2010).
Due to the value attributed to interventions, the availability in Spanish of
instruments is essential for clinical practice and research in Latin America; these
tools facilitate, through simple procedures and at low cost, an assessment of their
use in natural treatments within the most widespread theoretical frameworks. The
availability of instruments of this nature can contribute significantly to clinical
research and the training and supervision of novice therapists in the region.
In recent decades, a wide range of instruments that focus on the study of
interventions have been developed (for a comprehensive review see Hilsenroth,
Blagys, Ackerman, Bonge, & Blais, 2005). However, many of the measures
created are circumscribed to a specific type of approach or a specific manual,
having little use in the exploration of natural or integrationist treatments
(Hilsenroth et al., 2005). Furthermore, most instruments have developed a large
number of items, reducing their feasibility in large-scale studies (Hilsenroth et al.,
2005). Also, most of these instruments were developed in other languages, with
few adaptations to Spanish and the Latin American context. It is worth noting that
there have been endeavours carried out in this direction in the region: (1) the
Multi-Dimensional Classification of Psychotherapeutic Interventions (Roussos
et al., 2005), created in Argentina, studied four levels in the analysis of interven-
tions (descriptive, content, space-temporal and strategic) through external judges
and observation of clinical material; (2) Observation and Classification Guide of
Therapists Verbal Behaviour (Fernndez-lvarez et al., 2015), also developed in
Argentina, to analyse three types of verbalizations in therapists (questions, state-
ments and indications) and identify their personal style through independent
judges; (3) the Coding System of Therapeutic Activity (Valds, Tomicic, Prez,
& Krause, 2010) is a method of encoding speech through external judges,
developed in Chile, that analyses the content of the activity and the type of
communication actions during the therapy.
These three systems require recording clinical material and training coders;
this is sometimes not feasible as it multiplies the resources needed to conduct a
study. We believe alternative instruments in Spanish validated in Latin America
Spanish adaptation of the CPPS / Adaptacin al castellano de la CPPS 3

are necessary; tools that use simple but reliable procedures to evaluate the
therapeutic action and which use fewer resources, facilitating their incorporation
and complimenting existing systems.
Following that, this paper presents the adaptation to Spanish and analysis of
psychometric properties of the Comparative Psychotherapy Process Scale (CPPS,
Hilsenroth et al., 2005), an instrument that explores the degree of adherence to
Cognitive-Behavioural (CB sub-scale, which includes cognitive, cognitive-beha-
vioural and behavioural techniques) and Psychoanalytic-Interpersonal interventions
(PI sub-scale, which includes psychoanalytic, psychodynamic interpersonal and
interpersonal interventions) in non-manualized treatments. A recent study by De la
Parra (2013) on 640 clinicians, academics and students from more than 17 Latin
American countries found that most participants self-proclaimed belonging to psy-
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

choanalytic/psychodynamic (44.8%) or cognitive behavioural (31.9%) frameworks,


showing the relevance of both these frameworks in the region. Items in this scale
attempt to assess the distinctive characteristics of these two approaches by analysing
the extent to which a delivered therapy has used the technique prescribed by one
model and proscribed by others (Hilsenroth et al., 2005). Unlike other methods,
CPPS was not created to study adherence to a specific therapy or manual, but was
developed as a tool that can be applied to different natural treatments (Hilsenroth
et al., 2005). Finally, the CPPS is presented as a brief and simple scale (20 items on a
Likert scale) that can be completed by the patient, the therapist or external raters,
facilitating its application for conducting empirical studies on large samples.

Methodology
Two bilingual researchers translated the items from the original instrument (thera-
pist version) to Spanish independently. A third expert researcher selected the most
appropriate translation for each item, creating a first version of the instrument in
Spanish. A bilingual researcher conducted a back translation of that version (from
Spanish into English), which was compared with the original, without finding
substantive differences.
The Spanish version was then sent to four experts therapists (two cognitive-
behavioural and two psychoanalysts), asking them to rate on a scale of 0 to 6 the
extent to which each item was representative of their framework and provide
comments on its clarity. The experts scores on the scale showed evidence of
theoretical validity (Cognitive: CB = 5.65, PI = 2.4; Psychoanalysts: CB = 0.3,
PI = 4.7). Based on the comments from two experts, the item on the CB scale
The therapist interacts with the patient in a teacher-like (didactic) manner was
modified (see the final version of the scale as Appendix 1). Subsequent to that, the
items were adapted to build a patient version.

Participants
Through a non-probability sampling (snowball) a sample of 101 therapists who
self-labelled themselves as either cognitive or psychoanalytical therapists was
4 J.-M. Gmez-Penedo et al.

collected (Sample 1), and they were asked to complete the CPPS with data from
the last session they had with an adult patient. Sixty-seven percent were women,
with an average clinical experience of 16.16 years (SD = 13.40). Thirty therapists
were cognitive-behavioural, 49 psychodynamic-psychoanalytic, 20 cognitive-inte-
grative and two psychoanalytic-integrative. 57.4% of patients whose sessions
were studied were women, with an average age of 36.27 years (SD = 10.78).
15.8% were initiating treatment, 76.2% were at an intermediate stage and 7.9% in
their final stages. The diagnosis of patients according to their therapists was:
26.7% anxiety disorders, 17.8% mood disorders, 13.9% personality disorders,
9.9% neurosis, 9.9% borderline disorder, 3% life/personal development crisis and
18.8% other.
Through a non-probability snowball sampling a 15 therapist-patient dyad
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

sample was collected (Sample 2) where the patient and therapist had to complete
the CPPS. Therapists could be from other theoretical frameworks, beyond psycho-
analysis and cognitive-behaviourism. Of the therapists, 47% were women, with an
average clinical experience of eight years. Nine were cognitive, four were psycho-
analysts and two were from other frameworks.
Patients had an average age of 35.33 years (SD = 8.18), with 60% being
women. 26.7% were initiating treatment, 40% at an intermediate stage and 33.3%
in their final stages. The diagnoses of patients were: 40% anxiety disorders, 20%
life issues/personal development crisis, 13% personality disorders, 13% mood
disorder, 7% neurosis and 7% other.

Instruments
CPPS
The CPPS is a 20-item instrument (10 CB and 10 PI) presented in a Likert scale of
seven points (0 = Not characteristic, 6 = Extremely characteristic), which
assesses the extent to which cognitive-behavioural and psychodynamic-interper-
sonal interventions are representative of what occurred during a session
(Hilsenroth et al., 2005). For this study the therapist and patient version was used.
The original scale presented good psychometric properties (Hilsenroth et al.,
2005) with appropriate levels of inter-judge reliability and internal consistency
( = 9294; r corrected item-scales > .30). Furthermore, it showed concurrent
validity with regard to other scales and criterion validity, significantly differentiat-
ing psychodynamic from not psychodynamic treatments (Hilsenroth et al., 2005).

Procedures
For Sample 1, therapists were contacted through personal emails with information
about the objectives of research and asked to participate. Those therapists who
gave their consent were sent a link to an online page where the scale had been
uploaded. Therapists had to complete the form based on the last adult patient they
had attended. In turn, the therapists were asked to suggest other potential
participants.
Spanish adaptation of the CPPS / Adaptacin al castellano de la CPPS 5

The therapists from Sample 2 were given scales on a hard copy, the patient and
therapist version. The therapist had to select a patient and ask them if they wanted
to participate in the study. In the event they gave their informed consent, both
needed to complete the scale after finishing a specific session. Here, the therapists
are also asked to suggest new participants. In both samples the anonymity of the
subjects was respected.

Strategies for the analysis of data


For the analysis of the construct validity of the scale, an exploratory factor
analysis (EFA) was carried out using polychoric correlation matrix, suggested
for the analysis of ordinal variables (such as the Likert scales), and not Pearson
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

correlation matrices as is usual (Freiberg Hoffmann, Stover, de la Iglesia, &


Fernndez Liporace, 2013). Those matrices were also used to calculate
Cronbachs alpha ordinals (internal consistency) and corrected item-scale cor-
relations (homogeneity of the items). The polychoric correlations and analysis
derived were performed using the Software Factor 8 (Lorenzo Seva &
Ferrando, 2012).
Criterion validity was analysed using the SPSS 19, through two procedures:
(1) Logistic Regression seeking to predict the theoretical framework based on the
sub-scales CB and PI (predictive validity); (2) Spearmans Rho correlations (used
due to observed evidences of asymmetry and kurtosis) between the patients and
therapists scores (concurrent validity).
The construct validity, internal consistency and homogeneity of items were
analysed from Sample 1. Samples 1 and 2 were used to study the criterion
validity.

Results
Descriptive analysis (Sample 1)
Table 1 shows the mean and standard deviation of the items in Sample 1, together
with statistical asymmetry z(G1) and kurtosis z(G2). Scores lower than 1.96 or
above 1.96 allow us to reject the hypothesis of a symmetric and/or mesokurtic
distribution (Gonzlez, Abad, & Lvy, 2006). The hypothesis of a symmetric dis-
tribution is rejected in six items and a mesokurtic distribution is rejected in 11 items.

Construct validity (Sample 1)


The EFA used Parallel Analysis (PA) as the procedure to determine the number of
dimensions, suggesting a two-factor solution (Timmerman & Lorenzo-Seva,
2011). The Unweighted Least Squares (ULS) method was used as a method of
extracting factors, recommended for small samples (Freiberg Hoffmann et al.,
2013), with a Promin rotation (Lorenzo-Seva, 1999). Although the best ULS
method performance has been observed in dichotomous variables, the procedure
6 J.-M. Gmez-Penedo et al.

Table 1. Descriptive analysis of the CPPS items, therapist version (Sample 1).
Standard
Item Mean deviation zG1 zG2
(1) Exploration of uncomfortable feelings 5.20 1.45 3.92 2.16
(2) Explicit advice or direct suggestions 3.55 1.65 1.24 1.75
(3) Proposes subjects and activities. 3.57 1.86 0.27 2.62
(4) Connects feelings with past or current 4.80 1.30 1.19 0.28
perceptions.
(5) Focuses on similarities in relationships. 5.23 1.27 2.52 1.03
(6) Focuses on irrational or illogical beliefs 4.34 1.88 1.38 2.32
(7) Focuses dialogue on the therapeutic relationship. 3.52 1.65 0.96 1.55
(8) Encourages experimenting with and expressing 4.81 1.72 2.04 1.15
feelings in session.
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

(9) Specific activities or tasks external to the session. 4.05 2.42 0.31 3.44
(10) Addresses the avoidance of issues and changes 4.67 1.70 1.36 1.56
in mood.
(11) Explains criteria, techniques or treatment. 4.03 2.06 0.31 2.73
(12) Focuses on future situations. 3.71 1.53 0.24 2.06
(13) Suggests alternative forms of understanding. 4.75 1.45 2.14 0.90
(14) Identifies recurring patterns. 5.57 0.94 2.13 1.89
(15) Information on symptoms, disorders or 4.34 1.87 0.99 2.17
treatment.
(16) Allows the patient to start the dialogue. 5.58 1.32 2.50 1.21
(17) Suggests practice of behaviours learnt between 3.53 2.49 0.96 3.55
sessions
(18) Teaches techniques to deal with their 3.89 2.29 0.24 3.27
symptoms.
(19) Exploration of desires, fantasies, dreams or 3.66 1.94 1.16 2.37
memories.
(20) Interacts didactically. 3.42 1.98 0.87 2.53
CB scale 3.84 1.65 0.37 2.91
PI scale 4.78 0.80 0.23 0.15

has shown adequate performance in ordinal variables (Forero, Maydeu Olivares,


& Gallardo Pujol, 2009).
A suitable sample fit (KMO = .875; X2 = 1,375; gl = 190; p < .00001) and
Goodness of Fit Index (GFI = .99) was observed. The solution of two factors
explained 53.5% of the total variance. Rotated factor loadings showed a good fit
in regard to the original formulation, with all items presenting their main load in
their respective sub-scale (see Table 2). In turn all items except two had loads
higher than .4 in their sub-scales. Items 7 and 13, both psychodynamic, presented
loads of .347 and .359 respectively in the PI factor. Moreover, item 19 of the PI
factor, in addition to presenting a positive load on PI, presented a negative load of
similar magnitude in CB.
Spanish adaptation of the CPPS / Adaptacin al castellano de la CPPS 7

Table 2. Rotated factorial matrix solution of the CPPS, therapist version (Sample 2).
Item Scale CB-Load Factor PI-Load Factor Item-scale correlation
Item 1 IP .018 .611 .548
Item 2 CB .535 .066 .522
Item 3 CB .602 .130 .594
Item 4 IP .249 .544 .437
Item 5 IP .294 .625 .479
Item 6 CB .432 .352 .502
Item 7 IP .102 .347 .318
Item 8 IP .214 .584 .496
Item 9 CB .782 .125 .769
Item 10 IP .273 .445 .383
Item 11 CB .675 .150 .670
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

Item 12 CB .499 .127 .509


Item 13 IP .251 .359 .284
Item 14 IP .036 .592 .484
Item 15 CB .633 .190 .653
Item 16 IP .245 .634 .519
Item 17 CB .795 .060 .752
Item 18 CB .761 .140 .755
Item 19 IP .488 .425 .312
Item 20 CB .630 .107 .614
Note: Extraction method: Unweighted least squares (ULS); Method for determining the number of
factors: Parallel analysis (PA); Rotation: Promin.

Internal consistency (Sample 1)


Statistical Cronbach alpha ordinals were adequate both in the scale of interven-
tions CB (ordinal = .887) and in the PI (ordinal = .734).

Homogeneity (Sample 1)
All items except one presented item-scale corrected correlations superior to 3 (in
the range of .312.769; see Table 2). Item 13 (PI sub-scale) presented an item-
scale correlation of .284.

Criterion validity: predictive validity (Sample 1)


Using logistic regression, both sub-scales were able to predict whether the
therapists were psychoanalysts or cognitive (2 = 97.491, p < .001, degrees of
freedom = 2), showing a strong relationship between predictors and grouping
(Nagelkerke R2 = .826). Based on the model, 91.1% of therapists were classified
within their theoretical framework, both sub-scales contributing significantly (CB:
p < .001; PI: p = .027).
8 J.-M. Gmez-Penedo et al.

Criterion validity: concurrent validity (Sample 2)


The Spearmans Rho correlations between the scores of patients and therapists
were .435 (p < .001) in the total scale, .441 (p < .001) in PI and .422 (p < .001)
in CB.

Discussion
This research aimed to present the Argentine adaptation of the CPPS and perform
an initial analysis of its psychometric properties. The conducted EFA showed
evidence in favour of the construct validity of the scale, all items presenting their
main loads on their respective sub-scales and explaining 53.5% of the factors of
the total variance. However, some limitations in its factorial structure have been
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

observed. Items 7 and 13, of the PI subscale, presented loads below the threshold
of .4 (.347 and .359, respectively). However, they were not eliminated because of
the structure of the original scale and their theoretical importance in psycho-
analysis: they represent the focus on the patient-therapist relationship (item 7),
essential by the development of the transference, and the suggestion of alternative
understanding (item 13) as equivalent to the use of interpretations. These items
with low saturation factor must be assumed as factorially complex, with small but
non-zero loads that are close to the cut-off point (Ferrando & Anguiano-Carrasco,
2010; Lloret-Segura, Ferreres-Traver, Hernndez-Baeza, & Toms-Marco, 2014).
Item 19 (exploration of dreams, desires and fantasies), in addition to presenting a
positive load on its factor (PI), also had a negative load of an important magnitude
( .3) in the CB factor. Strictly speaking, this double load represents a threat to the
factorial stability of the scale. However, it should be noted that negative saturation
means that the item has an inverse relationship with regard to the items of this
sub-scale (CB), implying that the greater the exploration of dreams, desires and
fantasies, the less therapists use CB interventions, and vice versa. As both sub-
scales represent a prescribed set of interventions within a theoretical framework,
but also proscribed in the other, the result is expected and consistent with the
theory underlying the instrument.
This study also observed evidence of internal consistency and homogeneity of
the items. Both scales obtained values that exceeded .7. With the exception of
one item (13) all corrected item-scale correlations were greater than .3 and less
than .8, indicating an adequate level of homogeneity without the items being
redundant (Rattray & Jones, 2007). Item 13, which obtained a correlation of .284,
represents interpretation, an intervention that, despite being prototypical of psy-
choanalysis, is not recommended for use in conditions such as borderline dis-
orders in this framework (Grupo de trabajo OPD, 2006/2008). In psychoanalysis
there is a tendency to see borderline disorder, in a broader sense, as a personality
structure that encompasses different personality disorders and characteropathies
(Bernardi, 2010; PDM Task Force, 2006). Given this perspective, 23.8% of
patients from Sample 1 fall into the category of borderline disorders. This
significant percentage of participants may have undermined the association of
this item with other PI items.
Spanish adaptation of the CPPS / Adaptacin al castellano de la CPPS 9

Likewise, evidence of predictive validity was observed, reliably predicting,


using the CB and PI scores, the theoretical framework of the participants.
Correlations between scores of patients and therapists oscillated in the range
of .422.441 with a medium effect size (Cohen, 1992). This result would
indicate a partial evidence of concurrent validity of the CPPS. Considering
they measure the same construct, a stronger relationship between the two
versions would have been expected. An equivalent phenomenon was observed
in the validation of the original scale when comparing therapists with their
supervisors, with a Pearson rs in the range of .47.48 (Hilsenroth et al., 2005).
It could be stated that the complexity of clinical phenomena and asymmetric
weighting that the various participants (therapists, patients and supervisors)
perform from the various elements of therapy, favour this variability in
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

observations.
Comparing the results of this adaptation with the original scale is difficult
because this study used the therapist version and the original used independent
judges. Also, the original study of the scale does not perform factorial analysis.
Considering these caveats, the original scale had higher levels of reliability
( = .92.94; mean r corrected item-scale = .73) than this adaptation
( = .73.88; mean r corrected item-scale = .52). It is possible that the classifica-
tion of independent judges with extensive training in the CPPS increased the
internal consistency, compared to therapists who were unaware of the scale. In the
original scale, the association between therapists and their supervisors scores was
also explored, finding moderate correlation (r = .47.48), as occurred in this study
between therapists and patients scores (r = .42.44).
In a German adaptation of the instrument (using independent judges), high
levels of internal consistency (r corrected item-scale = .72.75) were also found,
along with evidence of criterion validity: the PI items were significantly higher in
psychoanalytic therapies and CB in cognitive therapies (Kallenbach, Will,
Kaufhold, & Leuzinger-Bohleber, 2015). Thompson-Brenner and Westen (2005)
used CPPS therapists forms but adapted them in 41 items for the treatment of
bulimia. Through an EFA they found three factors (PI, CB and complementary
treatments) explaining 44.9% of the total variance. Cronbachs alpha of factors
showed a greater variability than in this study ( = .67.91). Cognitive therapists
had significantly higher scores in CB and lower in PI than psychoanalysts. Other
studies exploring the psychometric properties of other adaptation of the instru-
ment have not been found.
In summary, the results of this research show the CPPS as a simple and useful
instrument for the study of psychoanalytic-interpersonal and cognitive-beha-
vioural interventions, with preliminary evidence of validity and reliability. As
the study of interventions is one of the main focuses of clinical research, an
instrument with these characteristics represents a valuable contribution to the
development of research on large samples, reducing the need to use many
resources. It can also help train therapists, by verifying adherence of novice
clinicians to the theoretical framework, serving for both therapists personal
records and for use in supervision and training.
10 J.-M. Gmez-Penedo et al.

Limitations and future research


The size of Sample 1 is an important limitation of this study. While for traditional
EFA studies using five subjects per item with a minimum of 100 subjects is
recommended (Gorsuch, 1983), criteria that this study complies with, current
authors such as Ferrando and Anguiano-Carrasco (2010) argue that at least 200
cases are necessary to ensure optimal conditions. This limitation may reduce the
accuracy of the analysis, such that the results of this study should be interpreted
with caution, as preliminary evidence of the validity and reliability of this
adaptation, whose properties we should continue exploring.
As for the size of Sample 2, correlations for studies where a large effect size is
expected, 28 subjects are recommended as an optimum sample number (Cohen,
1992). Having used a smaller sample reduces the statistical power of the analysis,
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

increasing the likelihood of false negatives (not observing existing correlations).


Therefore, the results from Sample 2 should also be interpreted with caution. Future
research should explore the association of CPPS therapist version with the patient
version in larger samples, as well as its association with the version of external
judges and with other interventions scales. Moreover, a confirmatory factor analysis
would provide further evidence to evaluate the factorial structure of the CPPS.
Some authors consider that beyond the specific interventions of each frame-
work, the common factors of the different therapies are the main agent of change
(Barber, 2009). Empirical research supports this view, showing that the use of
general interventions, based on strengthening or resolving ruptures in the ther-
apeutic alliance, favours change (Safran, Muran, & Eubanks-Carter, 2011). While
the CPPS does not have an exploration of common interventions, parsimonious
and simple instruments might be developed, along the same lines as this scale, to
assess common forms of intervention that fall under different frameworks.
Spanish adaptation of the CPPS / Adaptacin al castellano de la CPPS 11

Escala Comparativa del Proceso Psicoteraputico (CPPS):


anlisis de la validez y confiablidad de su adaptacin al
castellano

Existe una presuncin arraigada en clnicos, acadmicos e investigadores de que las


intervenciones, es decir las acciones verbales que los terapeutas realizan en sesin
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

dirigidas a sus pacientes (Roussos, Etchebarne, & Waizmann, 2005), son uno de los
principales mecanismos de cambio en psicoterapia (McCarthy & Barber, 2009).
Este supuesto ha encontrado sustento parcial mediante investigacin emprica en
distintos marcos tericos (Owen & Hilsenroth, 2011; Strunk et al., 2010).
Debido al valor atribuido a las intervenciones, es fundamental, para la prctica
e investigacin clnica en Latinoamrica, contar con instrumentos en castellano
que permitan, mediante procedimientos sencillos y a bajo costo, evaluar su
utilizacin en tratamientos naturales de los marcos tericos ms difundidos. La
disponibilidad de instrumentos de esta naturaleza puede contribuir significativa-
mente a la investigacin clnica, as como tambin al entrenamiento y supervisin
de terapeutas noveles en la regin.
En las ltimas dcadas se ha construido una gran diversidad de instrumentos
focalizados en el estudio de las intervenciones (para una revisin exhaustiva ver
Hilsenroth, Blagys, Ackerman, Bonge, & Blais, 2005). Sin embargo, muchas de
las medidas creadas estn circunscriptas a un tipo de abordaje o manual
especfico, teniendo escasa utilidad en la exploracin de tratamientos naturales o
integracionistas (Hilsenroth et al., 2005). Por otro lado, en su mayora los
instrumentos desarrollados cuentan con un gran nmero de tems, que disminuyen
la viabilidad en estudios a gran escala (Hilsenroth et al., 2005). Asimismo, la
mayor parte de estos instrumentos fueron desarrollados en otros idiomas, con-
tando con pocas adaptaciones al castellano y al contexto latinoamericano. En la
regin, cabe sealar tres esfuerzos en esta direccin: (1) La Clasificacin
Multidimensional de Intervenciones Psicoteraputicas (Roussos et al., 2005),
creada en Argentina, estudia cuatro niveles en el anlisis de las intervenciones
(descriptivo, contenido, temporo-espacial y estratgico) mediante jueces externos
y observacin de material clnico; (2) La Gua de Observacin del Terapeuta Aigl
verbal (Fernndez-lvarez et al., 2015) desarrollada, tambin en Argentina,
para analizar tres tipos de verbalizaciones en terapeutas (preguntas, afirmaciones e
indicaciones) e identificar su estilo personal, mediante jueces independientes; (3)
El Sistema de Codificacin de la Actividad Teraputica (Valds, Tomicic, Prez,
& Krause, 2010), un mtodo de codificacin del lenguaje mediante jueces exter-
nos, desarrollado en Chile, que analiza el contenido de la actividad y el tipo de
acciones de comunicacin en terapia.
12 J.-M. Gmez-Penedo et al.

Estos tres sistemas requieren de grabacin de material clnico y entrenamiento


de codificadores, lo que multiplica los recursos necesarios para realizar un
estudio, redundando en algunos casos en una merma de su viabilidad.
Consideramos que es necesario contar con instrumentos alternativos en castellano
validados en Latinoamrica, que mediante procedimientos sencillos pero confia-
bles permitan evaluar el accionar teraputico, con una menor utilizacin de
recursos, para que puedan ser incorporados y complementados con los sistemas
ya existentes.
En este contexto, este artculo presenta la adaptacin al castellano y un anlisis
de propiedades psicomtricas de la Escala Comparativa del Proceso
Psicoteraputico (CPPS, por sus siglas en ingls; Hilsenroth et al., 2005), un
instrumento que explora el grado de adherencia a intervenciones Cognitivo-
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

Conductuales (sub-escala CB; que incluye tcnicas cognitivas, cognitivo-conduc-


tuales y conductuales) y Psicoanalticas-Interpersonales (sub-escala PI; que
incluye intervenciones psicoanalticas, psicodinmicas interpersonales e interper-
sonales) en tratamientos no manualizados. En un estudio reciente de De la Parra
(2013) sobre 640 clnicos, acadmicos y estudiantes de ms de 17 pases latinoa-
mericanos, se observ que la mayora de los participantes se autoproclamaban
pertenecientes al marco psicoanaltico/psicodinmico (44.8%) o cognitivo-con-
ductual (31.9%), mostrando la relevancia de estos dos marcos tericos en la
regin. Los tems de esta escala pretenden evaluar las caractersticas distintivas
de estos dos enfoques analizando el grado en que cualquier terapia ha utilizado
tcnicas prescriptas por un modelo y proscripto por el otro (Hilsenroth et al.,
2005). A diferencia de otros mtodos, la CPPS no fue creada para estudiar la
adherencia a una terapia o manual especfico, sino que se desarroll como un
instrumento que puede ser aplicado a diferentes tratamientos naturales (Hilsenroth
et al., 2005). Finalmente, la CPPS se presenta como una escala breve y sencilla
(20 tems en una escala de tipo Likert) que puede ser completada por el paciente,
el terapeuta u observadores externos, facilitando su aplicacin para la realizacin
de estudios empricos en muestras grandes.

Mtodos
Dos investigadores bilinges tradujeron los tems del instrumento original
(versin terapeuta) al castellano de manera independiente. Un tercer investigador
experto seleccion la traduccin ms adecuada de cada tem, creando una primera
versin del instrumento en castellano. Una investigadora bilinge realiz una
traduccin cruzada de dicha versin (del castellano al ingls), que fue comparada
con la original, sin encontrar diferencias sustantivas.
La versin en castellano fue luego enviada a cuatro terapeutas expertos (dos
cognitivo-conductuales y dos psicoanalistas), solicitndoles que puntuaran en una
escala de 0 a 6 en qu medida cada tem era representativo de su marco y que
realizaran comentarios sobre su claridad. Los puntajes de los expertos mostraron
evidencias de validez terica en la escala (cognitivos: CB = 5.65, PI = 2.4;
psicoanalistas: CB = 0.3, PI = 4.7). En funcin de los comentarios de dos expertos
Spanish adaptation of the CPPS / Adaptacin al castellano de la CPPS 13

el tem de la escala CB El terapeuta interacta con el paciente como si fuese un


profesor (de manera didctica) fue modificado (ver la versin final de la escala en
el Apndice 1). Luego, los tems fueron adaptados para construir una versin
paciente.

Participantes
Mediante un muestreo no probabilstico (por bola de nieve) se convoc una
muestra de 101 terapeutas (Muestra 1) que deban auto-rotularse como cognitivos
o psicoanalticos, para completar la CPPS respecto de la ltima sesin que haban
tenido con un paciente adulto. El 67% eran mujeres, con una experiencia clnica
media de 16.16 aos (DS = 13.40). Treinta terapeutas eran cognitivos-conduc-
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

tuales, 49 psicoanalticos-psicodinmicos, 20 cognitivo-integrativos y dos psicoa-


nalitico-integrativos. El 57.4% de los pacientes cuyas sesiones fueron puntuadas
eran mujeres, con una edad media de 36.27 aos (DS = 10.78). El 15.8% se
encontraba iniciando el tratamiento, el 76.2% en una etapa intermedia y el 7.9%
en una etapa final. La distribucin por diagnstico de los pacientes segn los
terapeutas fue: 26.7% trastornos de ansiedad, 17.8% trastornos del estado de
nimo, 13.9% trastornos de personalidad, 9.9% neurosis, 9.9% trastorno border-
line, 3% crisis vitales/de desarrollo personal y 18.8% otros.
Mediante un muestreo no probabilstico por bola de nieve se convoc tambin
una muestra de 15 dadas de tratamientos (Muestra 2) en donde paciente y
terapeuta deban completar la CPPS. Los terapeutas podan ser de otros marcos
tericos, ms all del psicoanaltico y cognitivo-conductual. De los terapeutas, el
47% fueron mujeres, con una experiencia clnica media de ocho aos. Nueve eran
cognitivos, cuatro psicoanalistas y dos de otros marcos.
Los pacientes presentaron una edad media de 35.33 aos (DS = 8.18), siendo
el 60% mujeres. El 26.7% se encontraba en una etapa de iniciacin del trata-
miento, el 40% en una intermedia y el 33.3% en una etapa final. La distribucin
por diagnstico de los pacientes fue: 40% trastornos de ansiedad, 20% crisis
vitales/problemticas de desarrollo personal, 13% trastornos de personalidad,
13% trastornos del estado de nimo, 7% neurosis y 7% otros.

Instrumentos
CPPS
Es un instrumento de 20 tems (10 CB y 10 PI) presentados en una escala de tipo
Likert de siete puntos (0 = Nada caracterstico, 6 = Extremadamente
caracterstico), que evala en qu medida intervenciones cognitivo-conductuales
y psicodinmico-interpersonales, son representativas de lo acontecido en una
sesin (Hilsenroth et al., 2005). Para este estudio se utilizaron las escalas
versin terapeuta y paciente.
La escala original ha presentado buenas propiedades psicomtricas (Hilsenroth
et al., 2005) con adecuados niveles de confiabilidad inter-juez y consistencia
interna ( = .92.94; r tem-escalas corregidas > .30). Adems, ha mostrado
14 J.-M. Gmez-Penedo et al.

validez concurrente respecto de otras escalas y validez de criterio, diferenciando


significativamente tratamientos psicodinmicos de no psicodinmicos (Hilsenroth
et al., 2005).

Procedimientos
Para la Muestra 1 se contactaron a terapeutas mediante correos electrnicos
personales contndoles los objetivos de la investigacin y ofrecindoles participar.
A aquellos terapeutas que dieron su consentimiento, se les envi un link a una
pgina online en donde se encontraba la escala. Los terapeutas deban completarla
en base al ltimo paciente adulto que haban atendido. A su vez, se les peda a los
terapeutas que sugirieran otros potenciales participantes.
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

A los terapeutas de la Muestra 2 se les entregaron escalas en formato fsico


versin paciente y terapeuta. El terapeuta deba seleccionar un paciente y propo-
nerle participar en el estudio. En caso de que brindara su consentimiento infor-
mado, ambos deban completar la escala luego de terminar una sesin especfica.
Aqu tambin se les peda a los terapeutas la sugerencia de nuevos participantes.
En ambas muestras se mantuvo el anonimato de los sujetos.

Estrategias para el anlisis de los datos


Para el anlisis de la validez de constructo de la escala se realiz un anlisis
factorial exploratorio (AFE), usando matrices de correlacin policricas, sugeridas
para el anlisis de variables ordinales (como es el caso de las escalas de tipo
Likert), y no matrices de correlacin de Pearson como es habitual (Freiberg
Hoffmann, Stover, de la Iglesia, & Fernndez Liporace, 2013). Aquellas matrices
tambin se utilizaron para calcular los alfa de Cronbach ordinales (consistencia
interna) y las correlaciones tem-escala corregidas (homogeneidad de los tems).
Las correlaciones policricas y anlisis derivados se realizaron mediante el
Software FACTOR 8 (Lorenzo Seva & Ferrando, 2012).
La validez de criterio se analiz usando el SPSS 19, mediante dos procedi-
mientos: (1) regresin logstica buscando predecir el marco terico en base a las
sub-escalas CB y PI (validez predictiva); (2) correlaciones Rho de Spearman
(utilizadas por evidencias de asimetra y curtosis) entre los puntajes de los
pacientes y los terapeutas (validez concurrente).
La validez de constructo, consistencia interna y homogenidad de tems se
analiz a partir de la Muestra 1. Para estudiar la validez de criterio se utilizaron
las Muestras 1 y 2.

Resultados
Anlisis descriptivos (Muestra 1)
La Tabla 1 presenta la media y desvo estndar de los tems en la Muestra 1,
junto con estadsticos de asimetra z(G1) y curtosis z(G2). Puntajes inferiores a
1.96 o superiores a 1.96, permiten rechazar la hiptesis de una distribucin
Spanish adaptation of the CPPS / Adaptacin al castellano de la CPPS 15

Tabla 1. Anlisis descriptos de los tems de la CPPS versin terapeuta (Muestra 1).
Desvo
tem Media estndar zG1 zG2
(1) Exploracin de sentimientos incmodos 5.20 1.45 3.92 2.16
(2) Consejos explcitos o sugerencias directas 3.55 1.65 1.24 1.75
(3) Propone los temas y actividades. 3.57 1.86 0.27 2.62
(4) Conecta sentimientos o percepciones actuales 4.80 1.30 1.19 0.28
con pasadas.
(5) Centra la atencin en similitudes en las 5.23 1.27 2.52 1.03
relaciones.
(6) Focaliza en creencias irracionales o ilgicos 4.34 1.88 1.38 2.32
(7) Centra el dilogo en la relacin teraputica. 3.52 1.65 0.96 1.55
(8) Anima a experimentar y expresar sentimientos en 4.81 1.72 2.04 1.15
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

sesin.
(9) Actividades especficas o tareas fuera de la 4.05 2.42 0.31 3.44
sesin.
(10) Aborda la evitacin de temas y cambios en el 4.67 1.70 1.36 1.56
nimo.
(11) Explica los criterios, las tcnicas o el 4.03 2.06 0.31 2.73
tratamiento.
(12) Se centra en situaciones futuras. 3.71 1.53 0.24 2.06
(13) Sugiere formas alternativas de comprensin. 4.75 1.45 2.14 0.90
(14) Identifica patrones recurrentes. 5.57 0.94 2.13 1.89
(15) Informacin de sntomas, trastornos o del 4.34 1.87 0.99 2.17
tratamiento.
(16) Permite al paciente que inicie el dialogo. 5.58 1.32 2.50 1.21
(17) Sugiere prctica entre las sesiones de conductas 3.53 2.49 0.96 3.55
aprendidas
(18) Ensea tcnicas para afrontar sus sntomas. 3.89 2.29 0.24 3.27
(19) Exploracin de deseos fantasas, sueos o 3.66 1.94 1.16 2.37
recuerdos.
(20) Interacta de forma didctica. 3.42 1.98 0.87 2.53
Escala CB 3.84 1.65 0.37 2.91
Escala PI 4.78 0.80 0.23 0.15

simtrica y/o mesocrtica (Gonzlez, Abad, & Lvy, 2006). En seis tems se
rechaza la hiptesis de una distribucin simtrica y en once de una distribucin
mesocrtica.

Validez de constructo (Muestra 1)


En el AFE se utiliz el Anlisis Paralelo (PA) como procedimiento para determi-
nar la cantidad de dimensiones, sugiriendo una solucin de dos factores
(Timmerman & Lorenzo-Seva, 2011). Se utiliz el mtodo de Mnimos
Cuadrados no Ponderados (ULS) como mtodo para extraccin de factores,
recomendado para muestras pequeas (Freiberg Hoffmann et al., 2013), con una
rotacin Promin (Lorenzo-Seva, 1999). A pesar de que el mejor desempeo del
mtodo ULS se ha observado en variables dicotmicas, el procedimiento ha
16 J.-M. Gmez-Penedo et al.

mostrado un rendimiento adecuado en variables ordinales (Forero, Maydeu


Olivares, & Gallardo Pujol, 2009).
Se observ un adecuado ajuste muestral (KMO = .875; X2 = 1,375; gl = 190;
p < .00001) e ndice de Bondad de Ajuste (GFI = .99). La solucin de dos
factores explic un 53.5% de la varianza total. Las saturaciones factoriales rotadas
presentaron un buen ajuste en relacin a la formulacin original, con todos los
tems presentando sus cargas principales en su sub-escala respectiva (ver Tabla 2).
A su vez todos los tems menos dos presentaron cargas superiores a .4 en sus sub-
escalas. Los tems 7 y 13, ambos psicodinmicos, presentaron cargas de .347 y
.359 respectivamente en el factor PI. Por otra parte, el tem 19 del factor PI,
adems de presentar una carga positiva en PI, present una carga negativa de
similar magnitud en CB.
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

Consistencia interna (Muestra 1)


Los estadsticos alfa de Cronbach ordinales fueron adecuados tanto en la escala de
intervenciones CB ( ordinal = .887) como en la PI ( ordinal = .734).

Tabla 2. Matriz de solucin factorial rotada de la CPPS versin terapeuta (Muestra 2).
tem Escala Carga Factor-CB Carga Factor-PI Correlacin tem-escala
tem 1 PI .018 .611 .548
tem 2 CB .535 .066 .522
tem 3 CB .602 .130 .594
tem 4 PI .249 .544 .437
tem 5 PI .294 .625 .479
tem 6 CB .432 .352 .502
tem 7 PI .102 .347 .318
tem 8 PI .214 .584 .496
tem 9 CB .782 .125 .769
tem 10 PI .273 .445 .383
tem 11 CB .675 .150 .670
tem 12 CB .499 .127 .509
tem 13 PI .251 .359 .284
tem 14 PI .036 .592 .484
tem 15 CB .633 .190 .653
tem 16 PI .245 .634 .519
tem 17 CB .795 .060 .752
tem 18 CB .761 .140 .755
tem 19 PI .488 .425 .312
tem 20 CB .630 .107 .614
Nota: Mtodo de extraccin: Mnimos cuadrados no ponderados (ULS); Mtodo para determinar la
cantidad de factores: Anlisis paralelo (PA); Rotacin: Promin.
Spanish adaptation of the CPPS / Adaptacin al castellano de la CPPS 17

Homogeneidad (Muestra 1)
Todos los tems menos uno presentaron correlaciones tem-escala corregidas
superiores a .3 (en el rango .312.769; Ver Tabla 2). El tem 13 (sub-escala PI)
present una correlacin tem-escala de .284.

Validez de criterio: validez predictiva (Muestra 1)


Mediante una regresin logstica ambas sub-escalas permitieron predecir si los
terapeutas eran psicoanalistas o cognitivos (2 = 97.491, p < .001, grados de
libertad = 2), mostrando una relacin fuerte entre predictores y agrupamiento (R2
de Nagelkerke = .826). En base al modelo, el 91.1% de los terapeutas fueron
clasificados en su marco terico, contribuyendo significativamente ambas sub-
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

escalas (CB: p < .001; PI: p = .027).

Validez de criterio: validez concurrente (Muestra 2)


Los coeficientes Rho de Spearman entre los puntajes de los pacientes y los
terapeutas fueron de .435 (p < .001) en la escala total, .441 (p < .001) en PI y
.422 (p < .001) en CB.

Discusin
Esta investigacin tuvo por objetivo presentar la adaptacin argentina de la CPPS
y realizar una exploracin inicial de sus propiedades psicomtricas. El AFE
conducido mostr evidencias en favor de la validez de constructo de la escala,
presentando todos los tems sus cargas principales en sus sub-escalas correspon-
dientes y explicando los factores un 53.5% de la varianza total. Sin embargo, se
han observado algunas limitaciones en su estructura factorial. Los tems 7 y 13, de
la escala PI, presentaron cargas por debajo del umbral de .4 (.347 y .359,
respectivamente). No obstante, se decidi conservarlos para respetar la estructura
de la escala original y debido a su importancia terica en el psicoanlisis: ellos
representan la focalizacin en la relacin paciente-terapeuta (tem 7), fundamental
en la elaboracin de la transferencia, y la sugerencia de alternativas de
comprensin (tem 13), como equivalente al uso de interpretaciones. Estos tems
con baja saturacin factorial deben asumirse como factorialmente complejos, con
cargas pequeas pero no nulas y cercanas al punto de corte (Ferrando &
Anguiano-Carrasco, 2010; Lloret-Segura, Ferreres-Traver, Hernndez-Baeza, &
Toms-Marco, 2014). El tem 19 (exploracin de sueos, deseos y fantasas),
adems de presentar una carga positiva en su factor (PI), ha tenido tambin una
carga negativa de importante magnitud ( .3) en el factor CB. En un sentido
estricto, esta doble carga del tem representa una amenaza a la estabilidad factorial
de la escala. No obstante, cabe sealar que la saturacin negativa significa que el
tem presenta una relacin inversa respecto de los tems de dicha sub-escala (CB),
implicando que a mayor exploracin de sueos, deseos y fantasas, menor fue el
uso de intervenciones CB en los terapeutas, y viceversa. Al representar ambas
18 J.-M. Gmez-Penedo et al.

sub-escalas un conjunto de intervenciones prescriptas en un marco terico, pero a


la vez proscriptas en el otro, el resultado es esperable consistente con la teora que
subyace al instrumento.
En el trabajo se observaron tambin evidencias de consistencia interna y
homogeneidad de los tems. Ambas escalas obtuvieron valores de superiores a
.7. Con excepcin de un tem (el 13) todas las correlaciones tem-escala corregidas
fueron superiores a .3 e inferiores a .8, indicando un adecuado nivel de homo-
geneidad sin que los tems sean redundantes (Rattray & Jones, 2007). El tem 13,
que obtuvo una correlacin de .284, representa la interpretacin, intervencin que
si bien es prototpica del psicoanlisis, desde este marco no est recomendado su
uso en patologas como los trastornos borderline (Grupo de trabajo OPD, 2006/
2008). Desde el psicoanlisis existe una tendencia a ver al trastorno lmite, en un
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

sentido ms amplio, como una estructura de personalidad que engloba los dis-
tintos trastornos de personalidad y caracteropatas (Bernardi, 2010; PDM Task
Force, 2006). Teniendo en cuenta esta visin, el 23.8% de los pacientes partici-
pantes de la Muestra 1 entraran en la categora de trastornos borderline. Este
porcentaje importante de participantes puede haber menoscabado la asociacin de
este tem con el resto de los tems de PI.
Asimismo, se observaron evidencias de validez predictiva, pudiendo predecir
confiablemente, mediante los puntajes en CB y PI, el marco terico de los
participantes. Las correlaciones entre los puntajes de pacientes y terapeutas
oscilaron en el rango de .422.441, con un tamao de efecto medio (Cohen,
1992). Este resultado indicara una evidencia parcial de la validez concurrente de
la CPPS. Considerando que miden el mismo constructo, se esperara una
asociacin ms fuerte entre ambas versiones. En la validacin de la escala original
se observ un fenmeno equivalente, al comparar terapeutas con sus supervisores,
con rs de Pearson en el rango de .47.48 (Hilsenroth et al., 2005). Puede
interpretarse que la complejidad de los fenmenos clnicos y la ponderacin
asimtrica que los distintos participantes (terapeutas, pacientes y supervisores)
realizan de los diversos elementos de la terapia, favorecen esta variabilidad en las
observaciones.
Comparar los resultados de esta adaptacin con la escala original resulta
difcil, porque en este estudio se utiliz la versin terapeuta y en la original se
usaron jueces independientes. Asimismo, en el estudio de la escala original no se
realizaron anlisis factoriales. Teniendo en consideracin estas salvedades, la
escala original present mayores niveles de confiabilidad ( = .92.94; media r
tem-escala corregida = .73) que esta adaptacin ( = 73.88; media r tem-escala
corregida = .52). Es posible que la clasificacin de jueces independientes con un
entrenamiento exhaustivo en la CPPS haya aumentado la consistencia interna, en
comparacin con terapeutas que desconocan la escala. En la escala original
tambin se explor la asociacin entre puntajes de terapeutas y sus supervisores,
hallando correlaciones moderadas (r = .47.48), al igual que ocurri en este
estudio entre los puntajes de los terapeutas y los pacientes (r = .42.44).
En una adaptacin alemana del instrumento (usando jueces independientes),
se encontraron tambin niveles altos de consistencia interna (r tem-escala
Spanish adaptation of the CPPS / Adaptacin al castellano de la CPPS 19

corregida = .72.75), junto con evidencias de validez de criterio: los tems PI


fueron significativamente ms altos en terapias psicoanalticas y los CB en
cognitivas (Kallenbach, Will, Kaufhold, & Leuzinger-Bohleber, 2015).
Thompson-Brenner y Westen (2005), usaron la forma terapeutas de la CPPS
pero en una adaptacin para el tratamiento de bulimia de 41 tems. Mediante
un AFE encontraron tres factores (PI, CB y tratamientos complementarios)
explicando un 44.9% de la varianza total. Los alfa de cronbach de los factores
presentaron mayor variabilidad que en este estudio ( = .67.91). Los tera-
peutas cognitivos tuvieron puntajes significativamente mayores en CB y
menores en PI que los psicoanalistas. Al momento, no se han encontrado
ms estudios que exploren propiedades psicomtricas de otras adaptaciones
del instrumento.
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

En sntesis, los resultados de esta investigacin presentan a la CPPS como un


instrumento sencillo y til para el estudio de intervenciones psicoanalticas-inter-
personales y cognitivo-conductuales, con evidencias preliminares de validez y
confiabilidad. Siendo las intervenciones uno de los principales focos de
investigacin clnica, un instrumento con estas caractersticas representa un aporte
valioso para el desarrollo de investigaciones en grandes muestras, reduciendo los
recursos necesarios. Asimismo, puede ayudar al entrenamiento de terapeutas,
mediante el chequeo de la adherencia de clnicos noveles a su marco terico,
sirviendo tanto para registros personales de los terapeutas, como para utilizarse en
supervisiones y estructuras de formacin.

Limitaciones y futuras investigaciones


El tamao de la Muestra 1 es una importante limitacin de este trabajo. Si bien
para AFEs estudios clsicos recomiendan utilizar cinco sujetos por tem con un
mnimo de 100 sujetos (Gorsuch, 1983), criterio que cumple este trabajo, en la
actualidad autores como Ferrando y Anguiano-Carrasco (2010) sostienen que al
menos 200 casos son necesarios para asegurar condiciones ptimas. Esta
limitacin puede reducir la precisin de los anlisis, por lo cual los resultados
de este estudio deben ser interpretados con cautela, como una evidencia prelimi-
nar de la validez y confiabilidad de esta adaptacin, cuyas propiedades debern
continuar siendo exploradas.
En cuanto al tamao de la Muestra 2, para estudios correlaciones en donde se
espera un tamao de efecto grande, se recomiendan 28 sujetos como nmero
muestral ptimo (Cohen, 1992). El haber usado una muestra inferior, reduce el
poder estadstico de los anlisis, aumentado la probabilidad de falsos negativos
(no observar correlaciones existentes). Por ende, los resultados derivados de la
Muestra 2 deben tambin ser interpretados con cautela. Futuras investigaciones
deberan explorar la asociacin de la CPPS versin terapeuta con la versin
paciente en muestras ms grandes, as como su asociacin con la versin de
jueces externos y con otras escalas de intervenciones. Por otra parte, un anlisis
factorial confirmatorio aportara mayor evidencia para evaluar la estructura facto-
rial de la CPPS.
20 J.-M. Gmez-Penedo et al.

Algunos autores consideran que por sobre las intervenciones especficas de cada
marco, los factores comunes a las distintas terapias son el principal agente de cambio
(Barber, 2009). Investigaciones empricas sustentan esta postura, mostrando que el
uso de intervenciones generales, basadas en fortalecer o resolver rupturas en la
alianza teraputica favorecen el cambio (Safran, Muran, & Eubanks-Carter, 2011).
En tanto la CPPS no cuenta con una exploracin de intervenciones comunes,
deberan desarrollarse instrumentos parsimoniosos y sencillos, al estilo de esta
escala, para evaluar formas de intervencin comunes a distintos marcos.

Disclosure statement
No potential conflict of interest was reported by the authors / Los autores no han referido
ningn potencial conflicto de inters en relacin con este artculo.
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

References / Referencias
Barber, J. (2009). Toward a working through of some core conflicts in psychotherapy
research. Psychotherapy Research, 19, 112. doi:10.1080/10503300802609680
Bernardi, R. (2010). DSM-5, OPD-2 y PDM: Convergencias y divergencias entre los
nuevos sistemas diagnsticos psiquitrico y psicoanalticos. Revista de Psiquiatra del
Uruguay, 74, 179205.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155159. doi:10.1037/
0033-2909.112.1.155
De la Parra, G. (2013). Psychotherapy research in developing countries: The case of Latin
America. Psychotherapy Research, 23, 609623. doi:10.1080/10503307.2013.830794
Fernndez-lvarez, H., Castaeiras, C., Curtarelli, A., Garca, F., Gmez, B.,
Lichtenberger, A., & Corbella, S. (2015). Presentacin de una gua para la
observacin y clasificacin de la conducta verbal de los terapeutas. Terapia
Psicolgica, 33, 2334. doi:10.4067/S0718-48082015000100003
Ferrando, P. J. F., & Anguiano-Carrasco, C. (2010). El anlisis factorial como tcnica de
investigacin en psicologa. Papeles del Psiclogo, 31, 1833.
Forero, C. G., Maydeu Olivares, A., & Gallardo Pujol, D. (2009). Factor analysis with
ordinal indicators: A Monte Carlo study comparing DWLS and ULS estimation.
Structural Equation Modeling, 16, 625641. doi:10.1080/10705510903203573
Freiberg Hoffmann, A. F., Stover, J. B., de la Iglesia, G., & Fernndez Liporace, M.
(2013). Correlaciones policricas y tetracricas en estudios factoriales exploratorios y
confirmatorios. Ciencias Psicolgicas, 7, 151164.
Gonzlez, N., Abad, J., & Lvy, J. P. (2006). Normalidad y otros supuestos en anlisis de
covarianzas. In J. P. Lvy (Ed.), Modelizacin con estructuras de convarianza (pp.
3157). La Corua: Netbiblo.
Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum
Associates.
Grupo de trabajo OPD. (2006/2008). Seleccin del foco y planificacin de la terapia. En
Diagnstico psicodinmico operacionalizado (OPD-2). Manual para el diagnstico,
indicacin y planificacin de la psicoterapia. Barcelona: Herder.
Hilsenroth, M. J., Blagys, M. D., Ackerman, S. J., Bonge, D. R., & Blais, M. A. (2005).
Measuring Psychodynamic-Interpersonal and Cognitive-Behavioral Techniques:
Development of the Comparative Psychotherapy Process Scale. Psychotherapy:
Theory, Research, Practice, Training, 42, 340356. doi:10.1037/0033-3204.42.3.340
Kallenbach, L., Will, A., Kaufhold, J., & Leuzinger-Bohleber, M. (2015) The LAC
depression study - The effectiveness of psychoanalytic and cognitive-behavioral ther-
apy of chronic depressed patients: Measuring the therapeutic adherence using the
Spanish adaptation of the CPPS / Adaptacin al castellano de la CPPS 21

comparative psychotherapy process scale (CPPS). Paper presented at the 8th


European Conference on Psychotherapy Research, Klagenfurt.
Lloret-Segura, S., Ferreres-Traver, A., Hernndez-Baeza, A., & Toms-Marco, I. (2014).
El anlisis factorial exploratorio de los tems: Una gua prctica, revisada y actuali-
zada. Anales de psicologa, 30, 11511169. doi:10.6018/analesps.30.3.199361
Lorenzo Seva, U., & Ferrando, P. J. (2012). Manual of the Program FACTOR. Retrieved
from: http://psico.fcep.urv.es/utilitats/factor/
Lorenzo-Seva, U. (1999). Promin: A method for oblique factor rotation. Multivariate
Behavioral Research, 34, 347365. doi:10.1207/S15327906MBR3403_3
McCarthy, K. S., & Barber, J. P. (2009). The Multitheoretical List of Therapeutic
Interventions (MULTI): Initial report. Psychotherapy Research, 19, 96113.
doi:10.1080/10503300802524343
Owen, J., & Hilsenroth, M. J. (2011). Interaction between alliance and technique in
predicting patient outcome during psychodynamic psychotherapy. Journal of
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

Nervous & Mental Disease, 199, 384389. doi:10.1097/NMD.0b013e31821cd28a


PDM Task Force. (2006). Psichodynamic diagnostic manual. Silver Spring, MD: Allia.
Rattray, J., & Jones, M. C. (2007). Essential elements of questionnaire design and
development. Journal of Clinical Nursing, 16, 234243. doi:10.1111/jcn.2007.16.
issue-2
Roussos, A. J., Etchebarne, I., & Waizmann, V. (2005). Un esquema clasificatorio para las
wintervenciones en psicoterapia cognitiva y psicoanaltica. Anuario de Investigaciones
de la Facultad de Psicologa de la Universidad de Buenos Aires, XVIII, 5161.
Safran, J. D., Muran, J. C., & Eubanks-Carter, C. (2011). Repairing Alliance Ruptures.
Psychotherapy, 48, 8087. doi:10.1037/a0022140
Strunk, D. R., Brotman, M. A., & DeRubeis, R. J. (2010). The process of change in
cognitive therapy for depression: Predictors of early inter-session symptom gains.
Behaviour Research and Therapy, 48, 599606. doi:10.1016/j.brat.2010.03.011
Thompson-Brenner, H., & Westen, D. (2005). A Naturalistic Study of Psychotherapy for
Bulimia Nervosa, Part 2. Therapeutic Interventions in the Community. The Journal of
Nervous and Mental Disease, 193, 585595. doi:10.1097/01.nmd.0000178883.82580.18
Timmerman, M. E., & Lorenzo-Seva, U. (2011). Dimensionality Assessment of Ordered.
Polytomous Items with Parallel Analysis. Psychological Methods, 16, 209220.
doi:10.1037/a0023353
Valds, N., Tomicic, A., Prez, J. C., & Krause, M. (2010). Sistema de codificacin de la
actividad teraputica (scat-1.0): Dimensiones y categoras de las acciones comunica-
cionales de pacientes y psicoterapeutas. Revista Argentina de Clnica Psicolgica,
XIX, 117130.
22 J.-M. Gmez-Penedo et al.

Appendix 1 / Apndice 1
CPPS Versin Terapeuta y Juez externo

Consigna: Utilizando la escala presentada a continuacin, por favor punte cun


caracterstica es cada una de las siguientes afirmaciones respecto de las acciones
llevadas a cabo en la sesin. Por favor escriba el puntaje correspondiente en la lnea en
blanco que se encuentra al lado de cada tem.
Escala:
0 1 2 3 4 5 6
Nada caracterstica Poco Bastante Extrem adamente
caracterstica caracterstica caracterstica
Downloaded by [Juan Martin Gomez Penedo] at 23:15 29 July 2016

(1) El terapeuta alienta la exploracin de sentimientos percibidos como incmodos para


el paciente (e.g., enojo, envidia, excitacin, tristeza o alegra).
(2) El terapeuta da consejos explcitos o sugerencias directas al paciente.
(3) El terapeuta propone activamente los temas de discusin y actividades teraputicas.
(4) El terapeuta conecta sentimientos o percepciones actuales del paciente con
experiencias del pasado del paciente.
(5) El terapeuta centra la atencin en similitudes entre las relaciones del paciente que se
repiten en el tiempo, situaciones o personas.
(6) El terapeuta centra el dilogo en los sistemas de creencias irracionales o ilgicos del
paciente.
(7) El terapeuta centra el dilogo en la relacin entre el terapeuta y el paciente.
(8) El terapeuta anima al paciente a experimentar y expresar sentimientos en la sesin.
(9) El terapeuta sugiere actividades especficas o tareas (tareas para la casa) para que el
paciente realice fuera de la sesin.
(10) El terapeuta aborda la evitacin del paciente sobre temas importantes y cambios en
el estado de nimo.
(11) El terapeuta explica los criterios que existen tras sus tcnicas o su enfoque de
tratamiento.
(12) El terapeuta centra el dilogo en situaciones futuras de la vida del paciente.
(13) El terapeuta sugiere formas alternativas de comprensin de experiencias o
acontecimientos no reconocidos previamente por el paciente.
(14) El terapeuta identifica patrones recurrentes en las acciones, sentimientos y/o
experiencias del paciente.
(15) El terapeuta aporta datos e informacin al paciente respecto de sus sntomas, su
trastorno o del tratamiento.
(16) El terapeuta permite al paciente que inicie el dialogo a partir de aspectos, eventos o
experiencias significativas.
(17) El terapeuta sugiere explcitamente que el paciente practique entre las sesiones la(s)
conducta(s) aprendida(s) en terapia.
(18) El terapeuta le ensea al paciente tcnicas especficas para afrontar sus sntomas.
(19) El terapeuta alienta la exploracin de los deseos del paciente, fantasas, sueos o
recuerdos tempranos de la infancia (positivos o negativos).
(20) El terapeuta interacta con el paciente de forma didctica (como un docente).